1. Case Studies of Simultaneous Mining and Mineral Processing
Optimisation Applied to Platinum and Nickel Operations
By Steve Burks
Whittle Consulting
June 2012
Abstract
This paper develops the themes explained at the previous SAIMM Platinum conference.
Optimisation techniques can be used to significantly increase the value of mining businesses
by enabling better long term planning decisions. Open pit and underground mine design, mine
scheduling, cut-off grade and blending, stockpiling and the linking of these to flexible elements
of the metallurgical recovery processes are all evaluated together. Transport or sale of
intermediate products and the requirements of the product metal markets can also be
considered. Experience shows that net present value (NPV) can be increased by 5-35% or
more, usually even before the expenditure of significant amounts of project capital.
The case studies presented include an existing operation with multiple open pits and PGM
concentrators linked to offsite downstream processing facilities and with infrastructure
constraints for which the scheduling of current operations and the potential for expansion
were investigated; an existing operation with an integrated portfolio of mines, concentrators,
smelter and refinery for which the various possible sources of ore for processing were
prioritised to match the downstream process plant constraints and the product markets; an
underground operation in Zimbabwe for which the optimum mining width was calculated and
expansion plans were evaluated; and an underground mine with many separate mining
sections for which a revised schedule was developed aiming to keep the shaft system and
processing plant filled with the highest net value material available at any time throughout the
life of mine. In the concluding section of the paper, a few brief notes are included suggesting
how the strategic recommendations from Enterprise Optimisation studies may be
implemented by the Owner’s team.
Prepared for the SAIMM 5th
International Platinum Conference

2. . Page 2
Introduction
Many software packages exist to optimise various parts of a mining business in isolation.
However, it is rare for companies to optimise all parts of their operation or portfolio
simultaneously. In the past ten years, Whittle Consulting has focused on expanding the
boundaries of integrated optimisation, concentrating on issues faced by large and complex
mining and processing operations. Using advanced techniques an integrated geological,
mining, processing, transport and product model can be constructed. This is manipulated
mathematically to optimise the values of those variables that are considered negotiable.
Utilising this procedure, it is possible to develop long term plans that maximise the value of
large geological and technical plant asset portfolios. Improvements in net present value (NPV)
of the business in the range 5-35% have been demonstrated in many cases.
This methodology is very suitable to be applied to the PGM (platinum group metal) and related
sectors of the mining industry. Since the last paper was presented, Whittle Consulting has
completed several optimisation studies of PGM and Nickel operations or projects in a number
of different locations. In this paper, the author uses multiple case studies to illustrate how a
different suite of optimisation techniques in each case could be used to identify significant
overall value addition.
The paper concludes with a short section suggesting some of the typical action items needed
to implement the recommendations of this type of study.
A Reminder - What is Enterprise Optimisation?
Enterprise Optimisation can help to solve the production schedule challenges of mining and
processing operations with multiple pits, mining faces and underground mines, multiple metal
or mineral products, stockpiling and blending opportunities, and alternative processing
options. The combination of these features creates significant long term planning and
analytical problems and opportunities that often exceed the capabilities of commercially
available mining optimisation software.
Simultaneous optimisation aims to address all steps in the value chain and all assets in the
enterprise portfolio together, and does this while also considering all time periods of the
planned operation. This is a crucial additional complexity differentiating mining from other
businesses. An ore body is a depleting resource; when we decide what to mine and process
in one period, we constrain the available options for all future periods.
Figure 1 below illustrates the point. There is little or nothing that any enterprise can do to
improve the resource in the ground or the international market for the products, but all the
other steps illustrated in the figure can potentially be optimised.
Figure 1: The Generic Mining Enterprise Value Chain

3. . Page 3
Enterprise Optimisation concentrates on optimising the NPV of businesses. NPV is the sum
of discounted cash flows, normally calculated or forecast annually. It reflects the time value of
money and is considered to be a metric for planning and measuring the performance of any
business that will be understood and appreciated by executive management, shareholders
and other investors, and all stakeholders.
Philosophically, many mining businesses struggle to identify clear and consistent objectives.
For example, maximising metal production, maximising life of mine, minimising costs,
maximising resource recovery from the ground, maximising metal recovery from ore mined,
and maximising utilisation of equipment are all often cited as being key objectives for
operations. However, it is difficult to rank these against each other, and some of them conflict.
Enterprise Optimisation therefore focuses on a single objective, to maximise the economic
value of the business.
Application of Optimisation Principles to the Case Studies
The four case studies selected for this paper provided diverse optimisation challenges in
several links of the production value chain and therefore are used here to illustrate the flexible
nature of the mechanisms and techniques applied. In all cases, the names of the operation,
the parent mining company and the location have been withheld intentionally. Unfortunately
since the analyses in each case identified operational strategies which could affect the
valuation of the operation and all of the parent companies are publicly owned this is a sensitive
issue. Specific results could also not be provided in this paper which focuses on describing
the approach and mechanisms tested in each case and on the broad trends of the results.
In the case studies the optimisation work was applied to a portion or in one case the whole of
the typical value chain common throughout the industry. This is illustrated in Figure 2 below.
Figure 2: Alternative Process Configurations to be optimised In the PGM Secto
Case Study 1
New
Smelter
New
Mine
New
Concentrator
With DMS or ore sorting
Existing Mixed
Concentrator
Merensky
Concentrator
UG2
Concentrator
Platreef
Concentrator
Great Dyke
Concentrator
Conroast
Slag
Retreatment
And Disposal
Tailings
Retreatment
And Disposal Base Metal
Refinery
MC Plant
Precious
Metals
Refinery
Ni
(3 alternatives)
ACP
Existing
Smelter
Conc.
(PGM or
Chromite)
Matte
Cu
Co SO4
Platinum
Palladium
Rhodium
Ruthenium
Iridium
Au
Metallics
Third Party’s
Mine
Mine
MineMine
West Limb
MineMine
East Limb
Mine
MineMine
North Limb
MineMine
Zimbabwe
Stockpile
Stockpile
Matte

4. . Page 4
Description of the Task
This work was carried out for a large producer of nickel and platinum and related mainly to
one specific operating mine. This mine operates seven open pits all within approximately
fifteen kilometres of each other and which feed two concentrator plants. Open pit mining is
carried out using conventional blasting, load and haul methods and ore and waste are both
transported out of the pits using large dump trucks. The ore is crushed and can then be
transported and fed to either of the two milling and flotation plants. The flotation concentrate
containing PGM and base metals is transported to one of a number of possible smelters for
further processing.
The main objectives set for this Enterprise Optimisation study were:
 to facilitate deferral or minimisation of capital expenditure for the next 3 to 5 years
 to evaluate the merits of several alternative operating strategies. These included variable
versus capped mining rate, strategic stockpiling of low grade ore, deferral of waste
stripping where possible but with smaller short term pit phases to access higher grade
ore earlier, consideration of several concentrator upgrade or expansion strategies
 to determine the optimal mining rate matching the current concentrator capacity
 to establish the optimum dynamic mining and concentrator processing rates matching
current smelter and refinery capacity constraints
 to establish the optimal production rate taking into account the latest ore resource data
and applying current cost estimates for possible mine, concentrator and downstream
process plant expansions. The downstream plants receive material from other operations
and only have a certain amount of capacity available to process concentrate from this
operation. This had to be taken into account in the optimisation.
Methodology
The approach described here was also followed in the other case studies dealt with later so
will only be described once:
 An introductory meeting was held with the Owners’ operational and project team. All
disciplines likely to be stakeholders in the implementation of any recommendations later
were invited to be present. Consequently, there were personnel from the Geology, Mining,
Metallurgy, Transport, Procurement, Projects and Finance departments of the operation
itself as well as several equivalent representatives from Head Office. Optimisation
principles were discussed and explained, the current business plan of the operation was
presented and the personnel who would be responsible later for providing information
were identified.
 Resource model, operational, test work and study information was then collected from
each department at the operation and incorporated into a single Data Input or Business
Model to be used to prepare the input files for the Prober optimisation software. In addition

5. . Page 5
the current pit shell and phase designs were also transferred, to be used for the initial
stages of the optimisation.
 A Base Case optimisation was then prepared using Whittle Consulting’s proprietary
Prober software. This was compared with the operation’s current business plan and recent
operational results. Several of the input parameters were corrected in an iterative process
until the Prober results corresponded well with the current expectations of operational and
financial performance, the only difference being that the Prober Base Case utilised the
Scheduling optimisation mechanism and therefore advanced cash flow somewhat
compared with the mine’s plan by scheduling higher grade ore with low strip ratios earlier
if possible. The NPV of all other cases generated was compared to this Base Case. The
results were presented using a consistent format including Summary Tables, Dashboards
and Waterfall Charts to ensure that the Owner’s team could easily compare new results
with earlier cases.
 The next stage of work tested the six mining mechanisms of Enterprise Optimisation.
Dynamic cut-off grade optimisation and strategic stockpiling were added sequentially to
scheduling, all using the Owner’s optimised pit shells and phases. Then, the extensive
experience of Whittle Consulting’s specialists was applied using Gemcom Whittle Four X
software to prepare revised pit shells and phases with low initial stripping ratios, finer early
phases and quicker access to high grade ore pockets. The mining mechanisms were then
applied again sequentially based on the revised pit designs. After this, simultaneous
optimisation was applied with all of the mining mechanisms switched on. Dynamic mining
rate was tested by specifying the purchase cost, operational life and capacity of additional
mining equipment and using the Optimiser to determine how much additional mining
capacity would be optimal. This concluded Phase 1 of the contract.
 After approximately three months of activity, a typical Waterfall Graph was prepared, the
format of which is shown in Figure 3 below. This Graph was presented to the Owner’s
team and used as a basis to determine the direction of the next set of investigations.
Figure 3: Typical Waterfall Graph Showing the Effect of the Mining Mechanisms
Phase 2: Scenario Planning Based on the Current Planned Size of Operation
This operation supplies a significant percentage of the PGM and nickel produced by the parent
company in a sulphide concentrate that is then smelted and refined in offsite facilities. These

6. . Page 6
currently have a fixed capacity and due to mineralogical and other constraints the downstream
facilities have to limit the percentage of concentrate and metal they can accept in feed from
this mine. Phase 2 of the Enterprise Optimisation exercise worked within these limits and used
updated commercial parameters from the latest official business plan to determine the
optimum mining rates, concentrator schedules and processing practices for each year of the
projected life of mine. Specific scenarios considered included:
 Effect of dynamic annual mining rates on output of product metals
 Effect of varying blasting practices on concentrator throughput
 Effect of expanding concentrator capacity to the limit of water and power availability
 Possible divergent processing of hard and soft ore types
 Impact of changes in concentrate grade on metal recoveries.
In this phase of the work approximately 30 Prober Runs were completed over a six week
period, the output of each of which consisted of over 50 Tables and 70 Charts illustrating
trends in the key operating parameters as well as providing an overall financial analysis of the
effect of changes in the input settings. A few typical charts are illustrated in Figures 4 and 5
below, taken at random from two Runs completed for this case study.
Figure 4: Case Study 1 - Typical Production Charts from the Mining Operation

7. . Page 7
Figure 5: Case Study 1 - Typical Production Charts from the Concentrator Operations
Some common general principles of Enterprise Optimisation studies can be observed in
these charts:
 The annual mining rate is not constant. An optimised solution almost always includes
several years in which the maximum permitted mining rate is utilised in order to strip waste
from new pit phases in the process of being developed. Once a significant amount of ore
has been exposed in these phases then the optimum annual mining rate can often
decrease significantly for several years as the optimiser seeks to reduce variable mining
costs while simultaneously processing the highest value ore available in this period. This
typical mining rate trend can be seen clearly in Figure 4 above. In practice, this would
probably be achieved by having a fixed mining fleet operated by mine employees and
using contractors to achieve the additional tonnage needed in some years. Although the
operating cost to use the contractors is often higher than the mine’s own unit costs, this
approach reduces capital requirements and has less impact on mining fleet and driver
utilisation.
 At operations with multiple pits the optimiser may propose switching mining from one pit
to another for relatively short periods (ie several annual periods together) in order to
pursue the highest net value ore parcels while simultaneously complying with the
concentrator ore processing limits and the downstream metal processing constraints. This
would require a more complex planning strategy than is usually employed, and may
sometimes meet with reluctance from the operational mining team for good practical
reasons that have not been considered by the Prober software in its search for highest
NPV.

8. . Page 8
 Within the Prober software package, “stockpiling” implies that the ore will be processed
during the projected life of mine whereas “discard” applies to low grade mineralised
material which may in practice be set aside for later processing but which is not considered
sufficiently valuable to be processed in the optimised solution. Figure 4 illustrates both of
these principles being applied in Case Study 1. The chart showing stockpiled tonnage
implies that lower value ore is deferred by placing it on a strategic stockpile for several
years in order to permit higher grade ore to be mined and processed in the early years of
the schedule. The lower grade ore is then reclaimed from the stockpile and processed at
the end of the schedule. As long as the ore does not deteriorate while on the stockpile
leading to degradation of metallurgical performance later, this strategy usually adds value.
 Figures 4 and 5 also show cut-off grade optimisation being applied. Following the principle
of “It’s a Money Mine” the optimised solution aims to maximise the net cash flow from ore
being processed each year and therefore values ore by its monetary value from a basket
of product metals, taking into account the cost to process the ore as well as the intrinsic
value contained, rather than by simply considering the grade of any one specific metal.
Figure 5 shows that the major influence on value at this operation is the PGM content and
the average grade can be seen declining gradually during the period studied in order to
bring forward both net undiscounted cash and discounted cash flow as much as possible
to the early years of the operation. The charts showing base metal (nickel and copper)
grades do not exhibit the same trend, partly because of downstream base metal refinery
capacity constraints but mainly because the value of the base metals is subordinate to the
PGM content.
 Figure 5 illustrates the typical impact of constraints at various points in the value chain.
The optimiser specifies the ore to be sent to each concentrator and usually sends higher
grade ore to one location either to take advantage of better metallurgical performance (eg
higher recovery or greater specific processing rate for a particular ore type) or because
the location of the major mining activity in any period is closer to one of the concentrators,
or after consideration of both of these issues. It is also evident that the optimised solution
in these specific Runs was unable to keep both concentrators full in every year of the
schedule although this would intuitively seem to be necessary to maximise output and
therefore cash flow. In this instance, the principal constraint cannot have been the overall
mining rate limit preventing sufficient ore from being mined, since there was also
stockpiled ore available to be re-processed. The base metal grade trends offer an
explanation. It was not possible for more concentrate to be processed downstream due to
the high content of nickel and copper in these years exceeding the available base metal
refining capacity. The optimiser therefore reduced the amount of ore processed for several
years to reduce the concentrate production. This is an excellent illustration of the
interdependence of each step in the value chain.

9. . Page 9
Phase 3: Scenario Planning Including Possible Expansions
The final phase of work on this study utilised a modified resource block model provided by the
operation, investigated the likely impact of possible expansion of concentrator, smelter and
refinery processing facilities on the mining operation, considered different possible mining
durations and ensured that production was maintained between defined limits each year to
prevent major in-process inventory swings from year to year. It was completed in
approximately one month. The results of this portion of this work are speculative and sensitive
so are not illustrated in this paper.
In total the three phases of work on this study took over five months to complete, with over
seventy Prober Runs being completed. A cumulative cash flow chart from one of the interim
reports is produced below in Figure 6 with the monetary scale intentionally omitted. This
illustrates the typical profile of Enterprise Optimisation study results with significant cash flow
improvements in the first years and a much flatter profile later.
Figure 6: Case Study 1 – Typical Discounted Cash flow Comparison
Case Study 2
Description of the Scope
This work was carried out for an operation producing PGM and base metals in Zimbabwe from
an underground mine positioned on the Great Dyke. Referring to Figure 1, all of the Enterprise
Optimisation mechanisms other than the first two steps of pit optimisation can be applied to
any form of mineral mining and processing enterprise, not simply to those operations and
projects based on open pits.
The Great Dyke is a lopolithic geological structure extending for 550km across much of
Zimbabwe along a strike of about N10 E. It is between 4km and 11km wide and contains
layered ultramafic intrusions which support a number of economically viable PGM mines and
processing plants operated by several different companies. Because the mining operations

10. . Page 10
are normally shallow and grades are comparable with those across the Bushveld complex in
South Africa, the Great Dyke has been the focus of expansion and enterprise extension plans
for a number of mining companies in the past few years.
Most operations based on the Great Dyke mine a single reef in which the vertical thickness of
the main mineralised zone is normally about 2 to 3 metres. This reef usually extends across
the Great Dyke from East to West, outcropping on both sides of the Dyke and dipping to a
depth of several hundred metres in the centre of the E-W section. As can be seen in Figure
6, there is often a displacement of the peak values of base metals (nickel, copper and cobalt)
and platinum group metals (platinum, palladium, rhodium and gold constitute 3PGM+Au)
considered vertically through the reef. The point at which the platinum grade peaks is defined
in Figure 7 as the “Pt Marker”. The base metal distribution is skewed towards the hanging wall
and base metal value peaks 15-45cm above the platinum peak. The palladium peak, on the
other hand, occurs about 30cm below the platinum peak. There is a very rapid decline in PGM
grades above the Pt marker but a long tail of PGM value below the marker. Consequently the
typical mine on the Great Dyke is faced with two challenges: first, determining the optimum
stoping height (vertical thickness) and second deciding how far above the platinum peak to
mine in order to pursue base metal values.
Figure 7: Common Vertical Distribution of Metals in Great Dyke Ore Bodies
In Case Study 2, the Owner’s team therefore set a number of objectives for the Enterprise
Optimisation study within the overall context of maximising NPV:
 Determine correct stoping height for the current underground mine taking into account
different possible types of mechanised trackless mining equipment.

11. . Page 11
 Identify practicable improvements to the current mine plan and schedule to enhance cash
flow.
 Quantify financial benefits of possible modifications to the current metallurgical plant
operations.
 Quantify potential benefits of a proposed mine and plant expansion. If possible determine
optimum timing for this and confirm conceptually the long term commercial potential of an
expanded operation.
As with the other case studies work commenced with a kick off meeting, data collection and
preparation of a base case. For this study, collection of technical and cost parameters for the
underground mining operation was particularly important, but the whole list of data collected
is relevant and represents what is typically required:
 Five year operational and commercial business plan
 Operating cost data by activity and major mine and plant element
 Production statistics for July to November 2010
 Summary of stoping height (mining width) options with metal grade profiles for most
boreholes showing distribution of both proven and indicated material
 Process flow diagrams
 Plant process design criteria
 Five year forecasts of plant feed grade
 Five year mining production schedule
 Life of Mine production profile options including expansion possibilities
 Five year summary of strategic issues
 Summaries of mining challenges and key performance indicators
 Pillar designs and extraction factors for different mining widths and at different stope
levels (ie different depths below surface) of the current mining operation
 Study and test work reports dealing with possible recovery of PGM from tailings
 Five year “stay in business” capital schedule
 Copy of mine plan showing borehole locations, mined out areas and five year mining
area boundaries
 Mine ventilation plan and diagram
 Summary of major mining equipment availabilities
 Actual mining costs for a test section in comparison with the mining width being adhered
to elsewhere in the mine
 Effect of milling rate on recovery and grind on throughput capacity
 Effect of concentrator mass pull on recovery
 Planned expansion capacity
 Capacity limits at each step in the value chain
 Smelting and refining parameters
 Metal price forecasts

12. . Page 12
Qualitative Description of the Results of Case Study 2
Following the preparation, review and approval of a Base Case, the basic Data Input Model
was used to prepare Prober Runs for a range of stoping heights (constant mining width then
used for all levels of the mine in each case rather than trying to vary the stoping height at
different levels) and the usual range of tables and graphs were produced for each Run. A
Summary Table of results was also generated which was used to rank the options. This took
into account the possibility of needing to purchase additional low profile mining equipment for
some of the narrower mining widths and also considered a number of technical issues such
as the different extraction factors and support requirements needed if a decision was to be
made to change from current mining practices. Figure 8 below shows the headings of the
Summary Table. As with most operations, in Case Study 2 the Owners’ team wished to
consider other factors as well as NPV before reaching a decision. The Table facilitated this
and a decision was made to be recommended to the mine’s executive management team.
Figure 8: Case Study 2 – Decision Criteria Used to Select Stoping Height
There is less flexibility in an underground operation than there is in open pits and the basic
Lerchs-Grossman algorithm for maximising undiscounted cash flow and determining the outer
shell boundaries of an open pit cannot be applied to an underground mine. However, the Base
Case did still include scheduling and cut-off grade optimisation to a limited extent. The overall
shaft capacity limited production in this case and was therefore the major bottleneck inhibiting
further NPV improvements in the initial cases studied particularly as it would be difficult to
reject or stockpile low grade ore underground or separate it from higher grade ore before
sending it up the main shaft. However, individual production half levels and the development
stopes had some flexibility in capacity and there were definite opportunities for cheap de-
bottlenecking of some sections of the processing plant. Following determination of the
optimum mining width (this term is used synonymously with “stoping height”) investigations of
NPV for the current operational production capacity focused on:
 Strategic stockpiling of low grade ore on surface in cases where it could clearly be
identified (mainly from the development stopes)
 Improvement in concentrator recovery by increasing mass pull and sending more (lower
grade) concentrate to smelting and refining
 De-bottlenecking of the milling circuit leading to increased throughput in flotation and
additional output of metal in concentrate while simultaneously losing more PGM and base
metals to tailings.
Figures 9 to 11 compare the production profiles of three Prober Runs with similar input
parameters. For each set of three graphs, the same scales have been applied (ie the major

13. . Page 13
graduations on the y axis represent the same quantities) to make the comparisons as realistic
as possible without actually showing the projected production figures.
Figure 9 shows that Run 36 and Run 39 applied the same maximum annual ore feed capacities
to processing whereas Run 41 included de-bottlenecking of the milling circuit. For a small
outlay of capital it was expected that the plant would be able to process slightly more ore each
year without sacrificing recovery in the flotation circuit. Run 41 therefore showed increased
processing capacity in most years and also identified that the underground development plan
would be unable to match the ore requirement for processing in some years.
Figure 9: Case Study 2 – Effect of De-bottlenecking the Milling Circuit
Figure 10 shows that in Runs 39 and 41 a higher concentrate mass pull was used in the
optimisation model in comparison with Run 36. Run 39 produced more concentrate at a lower
grade than Run 36; Run 41 retained the lower grade and higher recovery of Run 39 but
produced even more concentrate in the early years of the schedule due to the faster milling
rate. Since the operation was processing a finite amount of ore the overall quantity of
concentrate generated in Run 41 over the whole life of mine was the same as in Run 39.

14. . Page 14
Figure 10: Case Study 2 – Effect of Higher Mass Pull and Increased Milling Rate on
Annual Concentrate Tonnage
The scale used in Figure 11 below has been manipulated to make it easier to see the
differences between the graphs and the vertical scale does not originate from zero. It illustrates
that both of these mechanisms advanced the rate of production of PGM and hence also of
cash. Run 39 generated slightly more PGM every year than Run 36 due to the higher mass
pull enhancing the recovery in the flotation circuit. This is a viable strategy for most PGM
producers as long as the downstream smelter capacity is sufficient to handle the extra
concentrate. Run 41 produced considerably more PGM in most of the first ten years of the
schedule after which production dipped sharply. The higher processing rate in this Run would
cause PGM recovery to drop and more metal to be discarded to the tailings dam but would
also enable faster production of PGM contained in increased amounts of concentrate.
The final portion of this Study looked at a potential expansion of current operations involving
development of two additional adjacent ore bodies and either expansion of the current plant
or construction of a new plant. The technical and cost data available to prepare these
scenarios for Prober were somewhat preliminary so none of the results can be reported here.
However, in principle the model was established to test how to prioritise the existing and new
mining operations and also to determine how large the new plant should be to produce the
optimum NPV.

15. . Page 15
Figure 11: Case Study 2 – Effect of Higher Mass Pull and Increased Milling Rate on
Annual PGM Production in Concentrate
Case Study 3
Description of the Scope
The third case study relates to a significant underground mining operation of a large nickel
and PGM producer. This operation extracts ore via a single large shaft but operates
approximately twenty half levels as discrete business areas (ie defined mining sections for
which production and cost reports are generated) and currently schedules mining from these
using a complex yet relatively unsophisticated spreadsheet based model aiming to produce a
blended mixture of ore from many locations each of which has different contained minerals
and diverse metal content. This enables the downstream concentrators to produce a
consistently suitable feedstock for smelting. At this operation, this is crucial to achieve
satisfactory metallurgical and cost performance in the smelter, avoid damage to the
refractories and ensure that environmental emissions (eg sulphur dioxide) remain within the
approved limits. The model is used for both short and long term planning. Historically it has
worked well but was facing challenges forecasting utilisation of the shaft hoisting capacity in
every year of the latest business plan which incorporated preliminary plans to increase hoisting
capacity.
Figure 12 illustrates possible mine structures in the PGM and nickel industries. The main shaft
location is shown near the centre of the diagram on the left with half levels identified on either
side. In the PGM industry this would typically be a plan view of an incline shaft with the half

16. . Page 16
levels proceeding down dip from top to bottom on the diagram. However, depending on the
size and shape of the ore body this principle could also be applied to a deeply dipping ore
body and a vertical shaft, for example as shown in the right hand diagram below. Ideally, the
operation would like to produce from all open sections of the mine at full capacity to keep unit
costs down and also fill the shaft hoisting capacity with the best ore available in order to
facilitate optimised cash generation. However, the more open mining areas there are, and the
more mineralogical or other constraints are imposed on the ore to be hoisted for processing,
the more difficult this scheduling challenge becomes.
Figure 12: Possible Progressive Mine Developments Utilising Many Mining Sections
In this case the scope of work was to assess the current life of mine plan by replicating it to
produce a Whittle Consulting Base Case, then to provide proof of concept for possible future
optimisation work by testing several of the standard Enterprise Optimisation mechanisms.
Qualitative Description of the Results of Case Study 3
The base case was developed using the existing financial model and life of mine (LOM) plan
including all twenty of the selected mining sections. Current and planned hoist limits were
applied as appropriate, then mine development time and cost, mining and processing
recoveries and costs for each section as well as downstream cost and revenue data applying
to ore processed from any section were applied to generate a life of mine schedule by section
and for the overall operation. The overall tonnage and grade profile for the mine is shown in
Figure 13 below. A declining grade profile can be seen clearly even in the mine’s business
plan on which this Base Case is closely modelled.
The next steps were:
 to permit flexing of the mining schedule to advance certain mining sections if NPV could
be improved

17. . Page 17
to permit flexing of the cut-off grades applied in each section. Some manual work was
done by the Owner’s team in the main production sections to develop multiple stope
designs with different cut-off grades. The optimiser was then used to select the optimum
cut-off grade for these sections before re-applying the schedule optimisation. Stockpiling
optimisation was not applied in this case study since the current practice is to process all
ore as it is hoisted
 to permit tonnage mined in individual sections to increase by 5%, still subject to the overall
hoist capacity limit.
Figure 13: Case Study 3 – Tonnage and Grade Profile for the Base Case
Figure 14 uses the same vertical scale as Figure 13 and shows the effect of these changes.
Tonnage hoisted achieves the maximum possible each year resulting in a slight decrease in
time to exhaust the defined ore body. Nickel grade peaks earlier and at a higher level than
previously. Both of these changes would result in cash flow being brought forward.
Figure 14: Case Study 3 – Tonnages and Grades After Applying Optimisation

18. . Page 18
Case Study 4
Description of the Scope
This was a portfolio optimisation carried out on behalf of a major producer of nickel with by
product PGM which was refreshed on a number of occasions to enable the Owner to update
the overall business plan for the combined operations. Since this work took place over several
years the description here is limited to a brief overview rather than a description of successive
phases of the work.
The company in question operated a large open pit and three underground mines at the time,
with the ore mined from these being supplemented by ore purchased from small operations
operated by others and potentially also from low grade stockpiles. The mining operations were
dispersed over a sufficiently large geographic area that it was essential for each main
operation to have its own sulphide ore concentrator. There were therefore three concentrators
in operation, each treating a variable blend of feedstocks and with different flow sheets,
metallurgical performance and cost structures. The concentrate from these plants was
transported to a central smelter and a refinery capable of producing final base metal products
and small quantities of PGM in intermediate products to be refined by others. The capacity of
the smelter exceeded that of the refinery necessitating some sale of smelter matte.
In addition to current operations the Owner was considering development of an additional
open pit mine and two new underground operations. These would mine and process a mix of
sulphide ore amenable to conventional metallurgical processing and some material for which
pressure leaching at the refinery was being considered as an alternative to smelting.
In summary, it was therefore possible to apply all ten of the Enterprise Optimisation
mechanisms to this portfolio of operations and projects, both at an individual operational level
and across the whole production and value chain.
The results illustrated in this report are from an intermediate stage of investigation and do not
necessarily represent the final outcome. Broadly, however, the objective was to prepare an
overall global optimisation for the entire operation, using input data from the latest block
models, pit and mine designs, process design criteria and capacity limits and commercial
forecasts all supplied by the Owner’s team. The following issues specifically needed to be
addressed in order to achieve the optimum solution:
 Tonnes and Nickel, Copper and Cobalt grade of matte, concentrate and mixed sulphide
to be fed to the refinery each year. In particular the refining process used required a
minimum copper:nickel ratio in the feed
 Tonnes of concentrate as well as Ni/Cu/Co grade, Ni:Cu ratio, sulphur content, sulphide
mineral ratios and total content, and Fe:MgO ratio in the total smelter feed and from each
concentrator, noting that the various concentrates potentially had very different chemical
and mineralogical compositions
 Talc content in ore to the concentrators
 Capacity of each concentrator and each mining operation
 Concentrate transport costs from each operation to the smelter
 Costs to produce each of the different possible types and grade of major product (eg nickel
metal, converter matte, mixed sulphide for sale)

19. . Page 19
 Application of grade banding, cut-off grade optimisation and strategic stockpiling at the
open pits, as well as application of the standard pit shell, phase and scheduling
mechanisms.
The overall solution in this case was influenced to the greatest extent by the refinery and
smelter constraints so in this case the trends illustrated below commence from the end of the
process and work back towards the mining operations.
Trends Observed in Case Study 4
Figure 15 shows an observation which could have encouraged more than one possible
outcome. The optimiser had been given the freedom to maximise revenue from the refinery
without many constraints. In this instance nickel was the greatest contributor revenue so the
selected schedule encouraged mining of ore and production of concentrate with the highest
possible content of nickel. The nickel refining capacity of the refinery was not infringed, but
the amount of cobalt fed to the circuit would have exceeded the cobalt removal and refining
capacity. The possible corrective action here could be to change the ore mix to reduce cobalt
in refinery feed, or alternatively to de-bottleneck the small cobalt processing circuit.
Figure 15: Case Study 4 – Refinery Feed Composition Constraints
As well as there being several possible nickel products for sale, the refinery could potentially
accept three different feedstocks: smelter matte, concentrate or mixed sulphide precipitated
from a low pressure leach process. The optimiser was able to determine the best mix of these
three feed materials each year of the life of mine schedule while still complying with the overall
capacity limit for the tonnage to be fed to the refining circuit. As can be seen in Figure 16
below, the blend proposed varied from year to year.
0
20
40
60
80
100
Refinery Nickel:Cobalt Ratio in Feed per Year
Minimum Ni/Co Limit Actual Ni/Co Ratio

20. . Page 20
Figure 16: Case Study 4 – Optimum Feed Blend to Nickel Refining
As mentioned earlier the portfolio included several mines and concentrators operated by the
Owner and could also buy in flotation concentrate from third parties. The grade, Ni:S ratio and
MgO content of the concentrate blend fed to the smelter needed to be controlled both for
technical reasons and in the interests of increasing cash flow. Figure 17 illustrates the
optimised concentrate production profile for one of the smaller mines. The concentrate from
this operation had a very low MgO content and MgO:Fe ratio whereas some of the other
concentrates had comparatively much higher MgO:Fe ratios. Since the smelter had a technical
requirement to keep MgO:Fe ratio in feed beneath a specified minimum, the optimum
schedule shown in Figure 17 below proposed reserving some of the concentrate from this
operation for several years before using it as smelter feed. This would probably require this
operation to be idled for several years which could clearly present some practical problems
for the operators. The net benefit would have to be measured against the operational issues
before taking a decision about this.
Figure 17: Case Study 4 –Control of Smelter Feed Composition
Annual Tonnes Feed to Refinery
Feed Type 3
Feed Type 2
Feed Type 1
Limit
Low MgO Content Concentrate Destination
Matte Refinery
Leach
Smelter

21. . Page 21
Figure 18 includes two charts showing how the optimisation process balanced mine
production with concentrator capacity. On the left hand side, the graph shows ore feed from
two open pits to a concentrator with a planned expansion in year 7 of the schedule. As can be
seen, the optimum tonnages from each pit to the concentrator varied considerably from year
to year and the ore mining capacity of the two pits matched or exceeded the plant capacity.
The right hand side chart, by contrast, shows a situation in which two underground mines were
unable to keep a shared concentrator fully supplied with ore feed. In practice there would
usually be an attempt to transfer ore from one concentrator to another but in this instance the
optimiser rejected this due to the distance between them and the ore transport costs.
Figure 18: Case Study 4 – Balancing Ore from Multiple Sources with Concentrator Capacity
Figure 19 shows the typical application of the stockpile optimisation mechanism. Higher value
ore (due to grade, mineralogy, strip ratio or a combination) from one of the open pits was
processed in the early part of the schedule with lower grade ore being reclaimed slowly later
at the rate consistent with available downstream processing capacity each year.
Figure 19: Case Study 4 – Use of Stockpiling to Increase Early Cash Flow
Conclusion – Implementation of Recommendations
Ore Feed to Concentrator from Two Pits
Pit 2 Pit 1 Limit
Pit 2 - Tonnes of ROM and Stockpiled Ore Processed
ROM Ex Stkp

22. . Page 22
The case studies in this paper tested the capabilities of Enterprise Optimisation in
many different ways and in each case some innovative and frankly surprising or counter
intuitive results were produced. The work was carried out for organisations of different levels
of complexity operating in diverse cultural environments. Notwithstanding this, Whittle
Consulting generally encounters similar issues each time a study is presented and the
conclusions need to be considered by the Owners’ team for possible implementation. The
potential barriers to implementation are quite numerous and have been described in earlier
papers. However, observation of the four case studies discussed here suggests that the
following generic approach to implementation could be reasonably successful in most cases:
 The potential implementation team needs to be involved in the initial framing of the scope
of work and objectives of the Optimisation study. This includes participation in the initial
kick off meeting.
 The Owner should appoint a single point of contact to manage data transfer, ensure
participation of a core team throughout the study and liaise with the operations
management team concerning the progress and the results. This person and other core
team members should attend an optimisation course or workshop to ensure that they
understand the basic principles and mechanisms.
 The participants in the Owner’s core team must be representative of all departments
throughout the full range of the operations and any expansion projects. This is crucial,
because decisions made by any single department cannot be made in isolation since they
have an impact on the performance and optimum operating conditions of other
departments. This also applies to operating divisions of a large enterprise.
 At the end of the study phase of the optimisation work the results need to be presented to
Executives and also the management team likely to be responsible for implementing the
recommendations, in a format that is familiar to the Owner and with potential benefits
clearly identified. The recommendations need to be in compliance with all constraints and
limits set by the Owner’s team and must be practicable to implement. The biggest potential
wins should be identified at this point and if possible prioritised for implementation first. If
these are successfully implemented and the gains can clearly be seen, further
implementation of lesser benefits will be much easier to motivate.
 The recommendations will in most cases not be based only on NPV. Although the sole
objective function for Enterprise Optimisation is maximisation of discounted cash flow this
does not mean that it is impossible or even difficult to produce tables and charts tracking
other metrics (eg cost per ounce of PGM produced) that may be of interest to the Owner’s
management team and which may be used by Executives from time to time to support
presentations to analysts and investors. These other stakeholders in the operation need
to be convinced of the benefits of the proposals.
 Having decided which recommendations may be appropriate to be adopted the Owner’s
team will usually need to put in place a change management exercise. This is likely to be
a challenging phase and is beyond the scope of this paper to discuss, but if done well it
could result in the great majority of the identified potential benefits being realised.